CHAPTER SIX Transmitter and Oscillator Systems 6.1 TRANSMITTER PARAMETERS A transmitter is an important subsystem in a wireless system. In any active wireless system, a signal will be generated and transmitted through an antenna. The signal’s generating system is called a transmitter. The specifications for a transmitter depend on the applications. For long-distance transmission, high power and low noise are important. For space or battery operating systems, high efficiency is essential. For communication systems, low noise and good stability are required. A transmitter can be combined with a receiver to form a transceiver. In this case, a duplexer is used to separate the transmitting and receiving signals. The duplexer could be a switch, a circulator, or a diplexer, as described in Chapter 4. A transmitter generally consists of an oscillator, a modulator, an upconverter, filters, and power amplifiers. A simple transmitter could have only an oscillator, and a complicated one would include a phase-locked oscillator or synthesizer and the above components. Figure 6.1 shows a typical transmitter block diagram. The information will modulate the oscillator through AM, FM, phase modulation (PM), or digital modulation. The output signal could be upconverted to a higher frequency. The power amplifiers are used to increase the output power before it is transmitted by an antenna. To have a low phase noise, the oscillator or local oscillator can be phase locked to a low-frequency crystal oscillator. The oscillator could also be replaced by a frequency synthesizer that derives its frequencies from an accurate high-stability crystal oscillator source. The following transmitter characteristics are of interest: 1. Power output and operating frequency: the output RF power level generated by a transmitter at a certain frequency or frequency range. 172 RF and Microwave Wireless Systems. Kai Chang Copyright # 2000 John Wiley & Sons, Inc. ISBNs: 0-471-35199-7 (Hardback); 0-471-22432-4 (Electronic) 2. Efficiency: the DC-to-RF conversion efficiency of the transmitter. 3. Power output variation: the output power level variation over the frequency range of operation. 4. Frequency tuning range: the frequency tuning range due to mechanical or electronic tuning. 5. Stability: the ability of an oscillator=transmitter to return to the original operating point after experiencing a slight thermal, electrical, or mechanical disturbance. 6. Circuit quality (Q) factor: the loaded and unloaded Q-factor of the oscilla- tor’s resonant circuit. 7. Noise: the AM, FM, and phase noise. Amplitude-modulated noise is the unwanted amplitude variation of the output signal, frequency-modulated noise is the unwanted frequency variations, and phase noise is the unwanted phase variations. 8. Spurious signals: output signals at frequencies other than the desired carrier. 9. Frequency variations: frequency jumping, pulling, and pushing. Frequency jumping is a discontinuous change in oscillator frequency due to nonlinea- rities in the device impedance. Frequency pulling is the change in oscillator frequency versus a specified load mismatch over 360  of phase variation. Frequency pushing is the change in oscillator frequency versus DC bias point variation. 10. Post-tuning drift: frequency and power drift of a steady-state oscillator due to heating of a solid-state device. Some of these characteristics can be found in an example given in Table 6.1. 6.2 TRANSMITTER NOISE Since the oscillator is a nonlinear device, the noise voltages and currents generated in an oscillator are modulating the signal produced by the oscillator. Figure 6.2 shows the ideal signal and the signal modulated by the noise. The noise can be classified as an AM noise, FM noise, and phase noise. Amplitude-modulated noise causes the amplitude variations of the output signal. Frequency-modulated or phase noise is indicated in Fig. 6.2b by the spreading of the FIGURE 6.1 Transmitter system. 6.2 TRANSMITTER NOISE 173 TABLE 6.1 Typical Commercial Voltage-Controlled Oscillator (VCO) Specifications Frequency ð f 0 Þ 35 GHz Power ðP 0 Þ 250 mW Bias pushing range (typical) 50 MHz=V Varactor tuning range Æ250 MHz Frequency drift over temperature À2 MHz=  C Power drop over temperature À0:03 dB=  C Q ext 800–1000 Harmonics level À200 dBc minimum Modulation bandwidth DC À 50 MHz Modulation sensitivity (MHz=V) 25–50 FM noise at 100-kHz offset À90 dBc=kHz or À120 dBc=Hz AM noise at 100-kHz offset À155 dBc=kHz or À185 dBc=Hz FIGURE 6.2 Ideal signal and noisy signal. 174 TRANSMITTER AND OSCILLATOR SYSTEMS frequency spectrum. A ratio of single-sideband noise power normalized in 1-Hz bandwidth to the carrier power is defined as lð f m Þ¼ noise power in 1-Hz bandwidth at f m offset from carrier carrier signal power ¼ N C ð6:1Þ As shown in Fig. 6.3, lð f m Þ is the difference of power between the carrier at f 0 and the noise at f 0 þ f m . The power is plotted in the decibel scale, and the unit of lð f m Þ is in decibels below the carrier power (dBc) per hertz. The FM noise is normally given as the number of decibels below carrier amplitude at a frequency f m that is offset from the carrier. Figure 6.4 shows a typical phase noise measurement from a Watkins–Johnson dielectric resonator oscillator (DRO) [1]. The phase noise is 70 dBc=Hz at 1 kHz offset from the carrier and 120 dBc=Hz at 100 KHz offset from the carrier. Here dBc=Hz means decibels below carrier over a bandwidth of 1 Hz. It should be mentioned that the bulk of oscillator noise close to the carrier is the phase or FM noise. The noise represents the phase jitter or the short-term stability of the oscillator. The oscillator power is not concentrated at a single frequency but is rather distributed around it. The spectral distributions on the opposite sides of the carrier are known as noise sidebands. To minimize the FM noise, one can use a high- FIGURE 6.3 Oscillator output power spectrum. This spectrum can be seen from the screen of a spectrum analyzer. 6.2 TRANSMITTER NOISE 175 Q resonant circuit, a low-noise active device, a phase-locked loop, or avoid the operation in a region of saturation. Many methods can be used to measure the FM or phase noise [2–4]. These methods include the spectrum analyzer method, the two-oscillator method, the single-oscillator method, the delay-line discriminator method, and the cavity discri- minator method. 6.3 FREQUENCY STABILITY AND SPURIOUS SIGNALS Slight electrical, thermal, or mechanical disturbances can cause an oscillator to change operating frequency. The disturbance may cause the oscillation to cease since it could change the device impedance such that the oscillating conditions described in Chapter 4 are no longer satisfied. Stability is a measure that describes an oscillator’s ability to return to its steady- state operating point. The temperature stability can be specified in three different ways. For example, at 10 GHz, an oscillator or a transmitter has the following temperature stability specifications: Æ10 KHz=  C, or Æ800 KHz over À30  Cto þ50  C, or Æ1 ppm=  C, where ppm stands for parts per million. At 10 GHz, Æ1 ppm=  C is equivalent to Æ10 KHz=  C. This can be seen from the following: Æ1 ppm=  C  10 GHz ¼Æ1  10 À6  10  10 9 Hz=  C ¼Æ10 KHz=  C A typical wireless communication system requires a stability range from 0.5 to 5 ppm=  C and a phase noise range from À80 to À120 dBc=Hz. FIGURE 6.4 Phase noise measurement for a WJ VC1001 DRO [1]. (Courtesy of Watkins- Johnson.) 176 TRANSMITTER AND OSCILLATOR SYSTEMS Frequency variations could be due to other problems such as frequency jumping, pulling, and pushing, as described in Section 6.1. Post-tuning drift can also change the desired operating frequency. The transmitter with good stability and low noise is important for wireless communication applications. To improve the stability, one can use (1) high-Q circuits to build the oscillators (examples are waveguide cavities, dielectric resona- tors, or superconducting resonators=cavities); (2) temperature compensation circuits; or (3) phase-locked oscillators or frequency synthesizers, which will be discussed later in this chapter. For an oscillator, spurious signals are the undesired signals at frequencies other than the desired oscillation signal. These include the harmonics and bias oscillations. The harmonic signals have frequencies that are integer multiples of the oscillating frequency. If the oscillating frequency is f 0 , the second harmonic is 2f 0 , and the third harmonic is 3f 0 , and so on. As shown in Fig. 6.5, the power levels of harmonics are generally well below the fundamental frequency power. A specification for harmonic power is given by the number of decibels below carrier. For example, second- harmonic output is À30 dBc and third-harmonic output is À60 dBc. For a compli- cated transmitter with upconverters and power amplifiers, many other spurious signals could exist at the output due to the nonlinearity of these components. The nonlinearity will cause two signals to generate many mixing and intermodulation products. 6.4 FREQUENCY TUNING, OUTPUT POWER, AND EFFICIENCY The oscillating frequency is determined by the resonant frequency of the overall oscillator circuit. At resonance, the total reactance (or susceptance) equals zero. Consider a simplified circuit shown in Fig. 6.6, where Z D is the active device impedance and Z C is the external circuit impedance. The oscillating (or resonant) frequency is the frequency such that ImðZ D ÞþImðZ C Þ¼0 ð6:2Þ FIGURE 6.5 Oscillating frequency and its harmonics. 6.4 FREQUENCY TUNING, OUTPUT POWER, AND EFFICIENCY 177 where Im stands for the imaginary part. The circuit impedance is a function of frequency only, and the device impedance is a function of frequency ð f Þ, bias current ðI 0 Þ, generated RF current ðI RF Þ, and temperature ðTÞ. Therefore, at the resonant frequency, we have Im½Z D ð f ; I 0 ; I RF ; T Þ þ Im½Z C ð f Þ ¼ 0 ð6:3Þ Electronic frequency tuning can be accomplished by bias tuning or varactor tuning. The bias tuning will change I 0 and thus change Z D , resulting in a new oscillating frequency. The varactor tuning (as shown in Fig. 6.7 as an example) will change CðV Þ and thus change Z C , resulting in a new oscillating frequency. The frequency tuning is useful for frequency modulation in radar or communication systems. For example, a 10-GHz voltage-controlled oscillator (VCO) could have a modulation sensitivity of 25 MHz=V and a tuning range of Æ100 MHz by varying the bias voltage to a varactor. For most systems, a constant output power is desirable. Power output could vary due to temperature, bias, frequency tuning, and environment. A specification for power variation can be written as 30 dBm Æ 0:5 dB, as an example. FIGURE 6.7 Varactor-tuned oscillator. FIGURE 6.6 Simplified oscillator circuit. 178 TRANSMITTER AND OSCILLATOR SYSTEMS A high-efficiency transmitter is required for space or battery operating systems. The DC-to-RF conversion efficiency is given by Z ¼ P RF P DC  100% ð6:4Þ where P RF is the generated RF power and P DC is the DC bias power. In general, solid-state transistors or FETs can generate power ranging from milliwatts to a few watts with an efficiency ranging from 10 to 50%. Solid-state Gunn diodes can produce similar output power at a much lower efficiency of 1–3%. IMPATT diodes can produce several watts at 5–20% efficiency at high microwave or millimeter-wave frequencies. For higher power, vacuum tubes such as traveling-wave tubes, Klystrons, or magnetrons can be used with efficiency ranging from 10 to 60%. Power-combining techniques can also be used to combine the power output from many low-power sources through chip-level, circuit-level, or spatial power combining [5]. In many cases, a high-power transmitter consists of a low-power oscillator followed by several stages of amplifiers. The first stage is called the driver amplifier, and the last stage is called the power amplifier. The power amplifier is normally one of the most expensive components in the system. Example 6.1 A 35-GHz Gunn oscillator has a frequency variation of Æ160 MHz over À40  Ctoþ40  C temperature range. The oscillator can be tuned from 34.5 to 35.5 GHz with a varactor bias voltage varied from 0.5 to 4.5 V. What are the frequency stability in ppm=per degree Celsius and the frequency modulation sensitivity in megahertz per volts? Solution Frequency stability ¼Æ160 MHz=80  C ¼Æ2MHz=  C ¼ A ðin ppm=  CÞÂ10 À6  35  10 9 Hz A ¼Æ57 ppm=  C Modulation sensitivity ¼ f 2 À f 1 V 2 À V 1 ¼ 33:5 À 34:5 GHz 4:5 À 0:5V ¼ 0:25 GHz= V ¼ 250 MHz=V j 6.4 FREQUENCY TUNING, OUTPUT POWER, AND EFFICIENCY 179 6.5 INTERMODULATION The intermodulation distortion and the third-order intercept point discussed in Chapter 5 for a receiver or mixer also apply to a power amplifier or upconverter in a transmitter. Figure 6.8 shows the curves for the fundamental and two-tone third- order intermodulation signals. Conventional high-power RF=microwave amplifiers were once used to handle only a single carrier communication channel. In this case, they could operate within the nonlinear region of the dynamic range without the risk of intermodulation products generation, thus avoiding channel interference. Currently, many of the contemporary communication systems operate in a multicarrier environment that allows an enhancement in bandwidth efficiency. They are very attractive whenever there is a large demand to accommodate many users within a limited spectrum but are required to operate with minimized adjacent out-of-band spectral emissions (spectral containment). These unwanted frequency components are primarily the result of intermodulation distortion (IMD) products produced by the multiple carriers propagating through nonlinear solid-state devices. Consider two signals f 1 and f 2 which are the input signals to a power amplifier, as shown in Fig. 6.9. The two signals will be amplified and the output power can be determined from the fundamental signal curve given in Fig. 6.8. The two-tone third- FIGURE 6.8 Nonlinear characteristics for a power amplifier. 180 TRANSMITTER AND OSCILLATOR SYSTEMS order intermodulation products ð2f 1 À f 2 and 2f 2 À f 1 Þ are also generated and appear in the output port. The power levels of these IM products can be found from the two- tone third-order intermodulation (IM3) curve given in Fig. 6.8. The IM3 power levels are normally well below the fundamental signals at f 1 and f 2 . If the frequency difference D is very small, the IM3 products are difficult to be filtered out, and it is important to keep their levels as low as possible. Other third-order distortion frequencies 3f 1 ,3f 2 ,2f 1 þ f 2 ,2f 2 þ f 1 , as well as the second-order distortion frequencies 2f 1 ,2f 2 , f 1 þ f 2 , f 1 À f 2 , are of little concern because they are not closely adjacent in frequency and they can be easily filtered out without any disturbance to the original signals f 1 and f 2 . In most wireless communications, one would like to have IM3 reduced to a level of less than À60 dBc (i.e., 60 dB or a million times below the fundamental signals). One way to reduce the IM3 levels is to use the feedforward amplifier concept. The amplifier configuration consists of a signal cancellation loop and a distortion error cancellation loop, as shown in Fig. 6.10 [6]. The signal cancellation loop is composed of five elements: an equal-split power divider, a main power amplifier, a main-signal sampler, a phase=amplitude controller, and a power combiner. This loop samples part of the distorted signal out from the main amplifier and combines it with a previously adjusted, distortion-free sample of the main signal; consequently, the main signal is canceled and the IM products prevail. The error cancellation loop is composed of three elements: a phase=amplitude controller, a linear error amplifier, and an error coupler acting as a power combiner. This loop takes the IM products from the signal cancellation loop, adjusts their phase, increases their amplitude, and combines them with the signals from the main power amplifier in the error coupler. As a result, the third-order tones are greatly reduced to a level of less than À60 dBc. Experimental results are shown in Figs. 6.11 and 6.12. Figure 6.11 shows the output of the main amplifier without the linearizer, where f 1 ¼ 2:165 GHz and f 2 ¼ 2:155 GHz. At these frequencies, IM 1 ¼ 2f 1 À f 2 ¼ FIGURE 6.9 Power amplifier and its IM3 products. 6.5 INTERMODULATION 181 [...]... over the same temperature range? FIGURE P6. 3 6.4 Calculate the reference signal frequency in gigahertz for the phase-locked system shown in Fig P6. 4 6.5 Determine the output frequencies of the frequency synthesizer shown in Fig P6. 5 for N1 ¼ 10 and N1 ¼ 20 Note that fR ¼ 1 GHz and N2 ¼ 100 PROBLEMS 193 FIGURE P6. 4 FIGURE P6. 5 6.6 In the synthesizer shown in Fig P6. 6, N1 ¼ 10, N2 ¼ 10, and fR ¼ 10 MHz... Fig P6. 7 provides 401 output frequencies equally spaced by 10 kHz The output frequencies are from 144 to 148 MHz FIGURE P6. 6 194 TRANSMITTER AND OSCILLATOR SYSTEMS The reference frequency is 10 KHz, and the local oscillator frequency is 100 MHz Calculate the minimum and maximum values for N FIGURE P6. 7 6.8 In Problem 6.7, if N ¼ 4600, what is the output frequency? 6.9 In the synthesizer shown in Fig P6. 9,... The resolution or increment in frequency is equal to the reference frequency fR To improve the resolution, the reference frequency can also be divided before it is connected to the phase detector This scheme is shown in Fig 6.20 A fixed frequency divider with division ratio of N2 is introduced between the crystal oscillator and the phase detector Zero output from the phase detector requires the following... 1.010 GHz The output spectrums are shown in Fig P6. 2 What are the frequencies for the IM3 products fIM1 and fIM2 ? If the output power levels for f1 and f2 signals are 192 TRANSMITTER AND OSCILLATOR SYSTEMS increased to 30 dBm, what are the power levels for fIM1 and fIM2 signals? Use f1 ¼ 1:010 GHz and f2 ¼ 1 GHz FIGURE P6. 2 6.3 A 10-GHz PLO is shown in Fig P6. 3 (a) Determine the reference frequency of... OSCILLATOR SYSTEMS FIGURE 6.18 IEEE.) A 57-GHz phase-locked source (From reference 8, with permission from Solution In a harmonic mixer, the RF signal is mixed with the multiple frequency of the LO to generate an IF signal The IF is given by fIF ¼ fRF À NfLO or fIF ¼ NfLO À fRF ð6:10Þ From Fig 6.18, the reference frequency is given by 57 GHz À 4  7  fR ¼ fR Therefore, fR ¼ 1:96551724 GHz The output frequency... Frequency synthesizers can be realized using a PLL and a programmable frequency divider, as shown in Fig 6.19 The signals applied to the phase detector are the reference signal from the crystal oscillator and f0 =N from the output of the frequency divider A large number of frequencies can be obtained by varying N , the division ratio As an example, if fR ¼ 1 MHz, we will have the output frequency ð f0... control voltage is used to vary the VCO frequency The process will continue until the VCO frequency (or phase) is aligned with the multiple of the crystal oscillator frequency The frequency divider is used to divide the output frequency of VCO by N to match the frequency of the reference oscillator Because of the tracking, the output of the PLL has phase noise characteristics similar to that of the... between the electrodes, mechanical forces will be exerted on the bound charges within the crystal, and an electromechanical system is formed that will vibrate at a resonant frequency The resonant frequency and the Q factor depend on the crystal’s dimensions and surface orientation The Q’s of several thousand to several hundred thousand and frequencies ranging from a few kilohertz to tens of megahertz are... þ10 dBm at frequencies of 1.8 and 1.810 GHz (Fig P6. 1) The IM3 power levels are À50 dBm The amplifier has a gain of 10 dB and an input 1-dB compression point of þ25 dBm What are the frequencies for the IM3 products fIM1 and fIM2 ? What are the power levels for fIM1 and fIM2 if the input power levels for f1 and f2 signals are increased to þ20 dBm? FIGURE P6. 1 6.2 A power amplifier has two input signals... frequency? 6.9 In the synthesizer shown in Fig P6. 9, if N3 ¼ 1000 and fR ¼ 1 MHz, what is the output frequency for N1 ¼ 100 and N2 ¼ 200? FIGURE P6. 9 REFERENCES 1 Watkin-Johnson Telecommunication Product Handbook, Palo Alto, CA., 1996, p 85 2 G D Vendelin, A M Pavio, and U L Rhode, Microwave Circuit Design, John Wiley & Sons, New York, 1990, Ch 6 3 I Bahl and P Bhartia, Microwave Solid State Circuit Design, . battery operating systems. The DC-to -RF conversion efficiency is given by Z ¼ P RF P DC  100% ð6:4Þ where P RF is the generated RF power and P DC is the DC bias. of frequency only, and the device impedance is a function of frequency ð f Þ, bias current ðI 0 Þ, generated RF current ðI RF Þ, and temperature ðTÞ. Therefore,